摘要
多数宏观经济变量时间序列有季节波动,如果季节波动是非线性的,采用经季节调整过的数据或传统季节模型等线性处理季节波动的方法可能就不再适用。本文基于季节时变平滑转换自回归(SEATV-STAR)模型,运用"特殊到一般"的非线性检验策略对我国工业增加值季度增长率季节波动进行研究。结果表明:(1)工业增加值的季节波动兼有结构时变和非线性改变,工业增加值的周期波动是线性的。(2)技术进步、体制变迁等因素使得工业增加值季节波动发生连续的结构时变,它们是季节波动变化的主要影响因素。(3)工业增加值周期波动对其季节波动有非对称影响;在工业增加值的波峰阶段,其季节波幅会减小,且1、2季度工业增长率有明显提高。
Most of the macroeconomic time series have seasonal fluctuations and if the seasonal fluctuations are truly nonlinear, then the methods of seasonal adjusted or traditional season models using linear methods to deal with seasonal fluctuations are may be improper. Based on SEATV-STAR model, ＂from the special to general＂ testing strategies ware applied to investigate the seasonal fluctuation of China＇s industrial added value and the results are as follows： seasonal fluctuation of industrial added value has the properties of structural time-varying and nonlinear change, while the periodic fluctuation is linear. The factors such as technological progress and economic system transition are the main influence factors to cause sea sonal fluetuation＇s continuous structural change. Besides, cyclical fluctuation of industrial added value results in seasonal fluctuation＇s asymmetric change. In the peak periods of industrial added value, seasonal amplitude reduce while quarter 1th＆2nd＇s growth speed improve significantly.
出处
《中国管理科学》
CSSCI
北大核心
2016年第4期10-18,共9页
Chinese Journal of Management Science
基金
国家社会科学基金资助重点项目(15ATJ003)
国家自然科学基金资助项目(71162017)
